A growing body of work documents average differences in confidence between men and women (Jakobsson 2012; Lundeberg, Fox, and Punccohaar 1994; Barber and Odean 2001). Some have argued that women's lower levels of confidence might explain why women are less likely to be promoted (Coffman 2014, Exley and Kessler 2019). This paper uses survey data from a select group of tenured economists at top US universities to answer whether women at the top of their careers are less confident than men and why. Analyzing economists' answers to questions about the economy and government policy, we document average differences in confidence along two margins. First, women are 7.3 percentage points less likely to give "extreme" answers in which they strongly agree or disagree with a statement. Second, women are less confident in the accuracy of their answer. Women express a level of confidence that is 0.22 points (or 0.09 standard deviations) lower than a comparable man, as measured on a scale of 1 (unconfident) to 10 (very confident). The results persist after controlling for the years since PhD award, the PhDawarding institution, the current institution, and the number of solo and -coauthored publications up to tenure. However, this confidence gap is largely driven by women being less confident when asked questions that are outside of their field of expertise. Women are as confident as men when answering questions related to their primary field but less confident when asked a question outside of their primary field. This result is not driven by men having a greater breadth of expertise than women. This suggests that -previously documented confidence gaps could be explained by women having a better sense of the bounds of their expertise than men.